-
Notifications
You must be signed in to change notification settings - Fork 323
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Alex/open3d ml tutorials #545
base: alex/open3d-ml_tutorials
Are you sure you want to change the base?
Alex/open3d ml tutorials #545
Conversation
+ Condense all notebooks into 3 notebook tutorials. + Replace SemanticKITTI with Toronto3D in training notebook (small size & fast training). + TODO: Pass through gramarly (may find multiple mistakes), please focus on content & structure.
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
@@ -0,0 +1,910 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You can access configuration values as object attributes (cfg.{property_name})
or dictionary key values (cfg['{property_name}']
)
Reply via ReviewNB
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done
@@ -0,0 +1,449 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@@ -0,0 +1,449 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Standard convention for object detection lidar datasets (e.g. KITTI / Waymo) is x(red) = forward, y (green) = left, z (blue) = up. Add XYZ labels to the axis.
Mention that datasets may instead of r (lidar reflectance) as feature, or no features at all.
A dataset is partitioned into train(ing), valid(ation) and test(ing) splits. Models use data from the training split to update network weights and performance is measured on the validation split. Hyperparameters can be adjusted to optimize performance on the validation split. The test split is never used during the training step and is only used to report the final performance.
Reply via ReviewNB
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Both PyTorch and TF can be installed. User code just needs to choose one inside a specific python file.
requirements-torch.txt
if Nvidia GPU is not available.
Reply via ReviewNB
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also talk about:
Continuing training from a checkpoint.
Point to the tensorboard tutorial for visualizing training progress.
Reply via ReviewNB
@@ -0,0 +1,910 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@@ -0,0 +1,583 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Re-write condensed tutorials based on suggestions.
Changes were made to 3 tutorial notebooks
Suggested changes were:
docs/tutorial/notebook/
for all data, model checkpoint logs etc. (all notebooks)01_config_files.ipynb
)01_config_files.ipynb
)02_datasets.ipynb
)02_datasets.ipynb
)Please note:
02_datasets.ipynb
will not run because the notebook does not have access to smaller version of semanticKITTI dataset.@sanskar107 can you share a downloadable link for that dataset that I canWill use shared link https://cdn.discordapp.com/attachments/903644960211992607/946792875423838288/semkitti.zip.wget
in the the notebook?Inference_on_a_custom_data.ipynb
,reading_a_config_file.ipynb
,reading_a_dataset.ipynb
,train_ss_model_using_pytorch.ipynb
andtrain_ss_model_using_tensorflow.ipynb
before merging because they are essentially duplicates. Keeping them here for reference temporarily.This change is